Update app.py

app.py

@@ -1,7 +1,38 @@
+import tensorflow as tf
+import numpy as np
 import gradio as gr
+from PIL import Image
 
-def greet(name):
-    return "Hello " + name + "!!"
+# Load the saved model
+model = tf.keras.models.load_model('cifar10_cnn_model.keras')
 
-demo = gr.Interface(fn=greet, inputs="text", outputs="text")
-demo.launch()
+# Define the CIFAR-10 class names
+class_names = ['airplane', 'automobile', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck']
+
+# Define a function to preprocess the input image
+def preprocess_image(image):
+    image = image.resize((32, 32))  # Resize image to 32x32
+    image = np.array(image) / 255.0  # Normalize pixel values
+    image = np.expand_dims(image, axis=0)  # Add batch dimension
+    return image
+
+# Define the prediction function
+def classify_image(image):
+    preprocessed_image = preprocess_image(image)
+    predictions = model.predict(preprocessed_image)
+    predicted_class = class_names[np.argmax(predictions)]
+    confidence = np.max(predictions)
+    return f"Prediction: {predicted_class} (Confidence: {confidence:.2f})"
+
+# Create the Gradio interface
+interface = gr.Interface(
+    fn=classify_image,
+    inputs=gr.Image(type="pil"),
+    outputs="text",
+    title="CIFAR-10 Image Classifier",
+    description="Upload an image of a CIFAR-10 category (e.g., airplane, cat, dog), and the model will classify it."
+)
+
+# Launch the app
+if __name__ == "__main__":
+    interface.launch()